Within the scope of complementary and alternative medicine (CAM), a number of exercise modalities — such as tai chi, qigong, yoga, and lesser-known movement therapies — have been studied as aiding persons with fibromyalgia. A meta-analysis of the current research literature found some evidence for the beneficial effects of these exercise types, with tai chi appearing to be most favorable.
Fibromyalgia affects an estimated 15 million persons in the United States alone, a majority are female, and diminished aerobic fitness and poor physical function in afflicted patients have been well documented. Exercise has been strongly recommended as an adjunct to medication therapy for fibromyalgia, but traditional forms of strenuous aerobic, stretching, and strength-building exercise often appear to have limited benefits for reducing pain while improving function, and patient discontinuation rates are sometimes unacceptably high.
In view of these concerns, Scott Mist and colleagues at the Oregon Health and Science University, Portland, conducted a systematic research review and meta-analysis of land-based CAM exercise therapies that have been increasing adopted by patients with fibromyalgia, including: qigong, tai chi, yoga, and several less familiar movement therapies [Mist et al. 2013]. For purposes of their review, the following definitions were used:
- Exercise was defined as “planned, structured physical activity whose goal is to improve one or more of the major components of fitness — aerobic capacity, strength, flexibility, or balance.” (Studies of exercise therapies conducted in water were excluded in this review.)
- CAM was defined as “a group of diverse medical and health care systems, practices, and products that are not generally considered to be part of conventional medicine.” (CAM approaches involving low levels of exertion, such as brief yoga with an emphasis on mindfulness or breathing, were excluded.)
Their search discovered 16 qualifying studies; 10 randomized controlled trials (RCTs) and 6 of a more observational nature in design. In total, 832 patients with fibromyalgia participated, with 490 allocated to CAM exercise interventions. The mean sample size in the studies was 52 (range 6–128 participants), and the median retention rate in all studies was 81%. The researchers noted that the studies were largely conducted in middle-aged women, but demographics summary data were not provided in the report. Overall, the strength of the included studies was rated as moderate-to-low.
The primary outcome endpoint of interest was FIQ total scores or FIQ pain scores, converted to standardized mean differences (ie, Cohen’s d) as measures of effect size [interpreting effect sizes was discussed in an UPDATE here]. The FIQ (Fibromyalgia Impact Questionnaire) is a widely used, self-administered, 20-question tool for assessing the current health (ie, physical function, pain, other symptoms) of patients with fibromyalgia in clinical and research settings [more info on FIQ here].
Writing in the March 2013 online edition of the Journal of Pain Research, Mist et al. report the following results [data adapted from tables in the report]:
- Six studies of tai chi demonstrated a pooled, large-sized effect = 1.14 (95% Confidence Interval, 0.88–1.39; P<0.001). Only 1 of the 6 studies did not achieve statistical significance and the remaining 5 were of good size with narrow confidence intervals, all of which suggest a reasonably strong level of evidence.
- There were 3 studies of yoga, with a pooled, moderate-sized effect = 0.45 (95% CI, 0.12–0.76; P=0.005). Only 2 of the 3 studies were statistically significant.
- Qigong was examined in 6 studies and there was a pooled, moderate-sized effect = 0.47 (95% CI, 0.25–0.69; P<0.001). However, only half of the individual studies achieved statistical significance and the outcomes overall varied extensively from each other, with wide confidence intervals (ie, the studies were highly heterogenous).
- Individual trials also were reported for the following modalities (with effect size, 95% CI, P-value): Pilates (0.77, 0.19–1.35, P=0.009); BMP, or Body Movement and Perception therapy (2.25, 1.67–2.84, P<0.001); Biodanza (3.92, 3.05–4.80, P<0.001). All 3 modalities exhibited large, statistically significant effect sizes, with relatively narrow confidence intervals; however, these were single trials of each modality needing replication for further confirmation.
NOTE: The individual trials itemized above add up to more than 16 in total because some studies examined more than one modality. Pilates involves nonimpact strength, flexibility, and breathing exercises. Body Movement and Perception (or, Rességuier) is based on selected low-impact gymnastic movements integrated with postural exercises. Biodanza, or “life dance,” most often uses dance and related movements to optimize self-development and deepen self-awareness.
According to the researchers, only 2 participants reported any adverse effects (increased shoulder pain and plantar fasciitis), and none of the studies found any serious adverse events. Therefore, given the lack of negative effects, and the medium-to-high effects sizes for pain reduction and other benefits, Mist and colleagues state, “there is little risk in recommending these modalities as a critical component in a multimodal treatment plan, which is often required for fibromyalgia management.”
COMMENTARY: Systematic reviews and meta-analyses can be robust methods for establishing the quality and strength of evidence for pain management therapies. An understanding of these statistical methods was discussed in an UPDATE article [here].
In this meta-analysis by Mist et al., the modality with the best data profile and largest effect size — suggesting a higher level of strongly favorable evidence — was tai chi. Second choice would be qigong (albeit the data were somewhat inconsistent), and yoga would be third. The other 3 modalities appear to have promise, but they were solitary trials, so the reliability and strength of evidence cannot be determined. [Note: the Mist et al. article is open access, allowing interested readers to download and examine the forest plots and other data for themselves. See link in the reference below.]
The generally small sample sizes in the included studies (mean n=52) and predominance of middle-aged female subjects limits the statistical power and generalizability of study outcomes to other clinical populations. Another limitation, noted by the researchers, was that only a single interventionist (ie, exercise instructor) was involved in each of the trials. It is possible that a skilled, caring, and/or charismatic instructor in some trials, but not others, might have influenced beneficial outcomes.
The researchers write that all but 2 studies report positive outcomes; however, the study outcomes data indicate that only 1 of the studies was negative (ie, effect size = –0.42 in a study of qigong). Furthermore, while the rest of the data point-estimates were indeed positive (ie, point effect size estimates >0.0) a third of them (5/15) were statistically non-significant, which might temper enthusiasm regarding the strength of the evidence overall.
Prior Pain-Topics UPDATES have discussed exercise modalities — including tai chi and, most commonly, yoga — as benefitting various acute and chronic pain conditions [series here]. More specifically, a study by Wang et al. reported in 2010 [UPDATE here] found that tai chi afforded significant relief from muscle pain, better sleep, and a higher quality of life with less depression in persons with fibromyalgia (this study also was included in the meta-analysis by Mist et al.).
It is of some concern that most of the trials in the meta-analysis by Mist and colleagues, no matter how small the number of subjects, used multiple assessment instruments for detecting numerous endpoints. This increases the chances that at least some positive statistically significant results might be produced; however, it also increases the risk of Type I error, or finding significant results due to random chance alone (ie, false positives). Mist et al. were wise to focus only on one measure — FIQ scores — even though this might have limited the scope of potentially beneficial therapeutic effects that could be assessed and discussed in their report.
Meta-analyses in most areas of pain research also highlight the problem of studies examining the same topic using different patient-selection criteria, disparate outcome measures, and other methodological variations. This sort of inconsistency challenges the validity of combing data from the different studies in aggregated, or pooled, estimates of effect size. In fact, Mist et al. noted that there “was a significant amount of heterogeneity” across the studies they examined [particularly evident in the analysis of qigong] and, while they did not report calculating tests of heterogeneity, they did appropriately use random-effects modeling in their pooling of data.
As often is the case, Mist and colleagues advise that there is a need for large, meticulously designed and rigorously executed trials with active parallel arms — such as comparing traditional aerobic exercise with a CAM-oriented approach — to extend this body of evidence. However, it could be a many years before such studies are conducted and, then, it is questionable whether they will use methodologies that are statistically comparable to earlier trials. Clinical research is sometimes a very “messy business.”
REFERENCE: Mist SD, Firestone KA, Jones KD. Complementary and alternative exercise for fibromyalgia: a meta-analysis. J Pain Res. 2013;6:247-260 [available here].
Don’t Miss Out. Stay Up-to-Date on Pain-Topics UPDATES!
Register [here] to receive a once-weekly e-Notification of new postings.